Tapping into existing household survey data for research or policy use: hands-on exercise on water access in Kinshasa[Exploiter les données d’enquêtes ménages pour la recherche ou la décision publique : guide et étude de cas sur l’accès à l’eau à Kinshasa]

In most countries, national statistical offices periodically run large surveys that provide outstanding insights on several subjects: social, economic, health, cultural, political. In many cases, this data is only used to produce nationally aggregated indicators that feed international statistical portals (WDGs, SDGs, World Bank, WHO, United Nations…). However, it is possible to do much more with the raw data collected during these surveys: calculate other indicators, cross different variables, run analysis for subnational areas, as well as more sophisticated analysis. This survey data is accessible to scholars, students, policy makers or practitioners and is most useful for the appraisal, monitoring and evaluation of development projects, programs and public policies. We first describe the typology of data generally available in developing countries, its possible uses and how to obtain it. Then we provide detailed guidelines on how to analyze it, using the survey package, available for R, a free and open source statistical software. We illustrate the step by step procedure for survey data processing using Democratic Republic of the Congo as an example: several surveys of different type and from different dates (MICS 2010, 1-2-3 2012 and DHS 2014) are analyzed to understand the levels and trends of access to drinking water, in particular in the Kinshasa megacity.